Why Your Ads Work for a Week, Then Stop

It is one of the most common complaints we hear about paid social. The first week was excellent. Enquiries came in, the cost per lead looked sensible, you started planning where to put the extra business. By day ten the same ad, at the same budget, to the same audience, brings in half as much. By day twenty it brings in almost nothing, and the cost per lead has doubled.

Nothing broke. No one changed a setting. The ad did not get worse. What changed is the relationship between three things: how many people you can reach, how many times you have already reached them, and how interesting your ad still is on the fourth viewing. Those are two distinct problems — creative fatigue and audience saturation — and they need different fixes. Most owners are sold a fix for one when they have the other.

What "stopping" actually looks like in the report

Before you diagnose anything, look at two numbers over time, week by week, not as a single lifetime average.

Frequency. This is simply impressions divided by reach — the average number of times each person saw your ad. Meta reports it directly. If your frequency for the last seven days is 1.4, most people have seen the ad once. If it is 6.0, you are shouting at the same crowd six times a week.

Click-through rate by week. Break the campaign into seven-day chunks and compare. If your click-through rate — the share of people who saw the ad and clicked — falls steadily while frequency climbs, you have a delivery problem, not a landing page problem.

The pattern matters more than any single figure. Falling clicks with rising frequency is fatigue or saturation. Steady clicks with falling enquiries is something after the click: your landing page or form, your stock, your phone going unanswered. Those are completely different repairs, and swapping creative will not help the second one at all.

The arithmetic of a small audience

Saturation is mostly arithmetic, and the arithmetic is unforgiving for local businesses. Work through a typical case.

Say you sell modular kitchens and you target adults aged 30 to 55 within 10 km of Laxmi Nagar, with an interest filter or two. Meta estimates your audience at, let us say, three lakh people. That already sounds large. But an estimated audience size is not the number of people the platform can actually put your ad in front of cheaply — it is the pool it will draw from, and it draws from the cheapest, most responsive slice first.

Now spend ₹1,500 a day. If your cost per thousand impressions works out to ₹200 — and it varies enormously by placement, season and competition, so check your own account rather than trusting that figure — you are buying 7,500 impressions a day. Over thirty days that is 2.25 lakh impressions against a three lakh pool.

Those impressions do not spread evenly. The delivery system finds people who respond and shows them the ad again, because that is what you asked it to optimise for. So in practice you might be hitting perhaps 40,000 people five or six times each, rather than 2.25 lakh people once. The responsive slice gets used up. Everyone in it who was going to enquire has enquired. What is left is people who have decided, repeatedly, not to click.

Then the cost rises, and the rise is not a punishment. You are simply now bidding to reach people who are less interested, and the auction prices that accordingly.

The narrower your geography, the faster this happens. A national D2C brand can run one ad for months. A dentist in Preet Vihar can exhaust a realistic catchment in ten days at ₹2,000 a day.

Fatigue and saturation are different problems

Creative fatigue means the same people have seen this specific ad enough times that it no longer registers. The offer is still relevant; the execution has gone invisible. The tell: a new creative to the same audience recovers performance. Cost per result drops again immediately.

Audience saturation means you have run out of suitable people, not out of ideas. The tell: a genuinely new creative to the same audience does nothing. Performance stays flat. Only changing who you are talking to moves it.

This single test — new creative, same audience — is the cheapest diagnostic available to you. Run it before you change anything structural. It costs a few days of budget and it tells you which of the two doors to walk through.

Symptom	Likely creative fatigue	Likely saturation
Frequency	Rising, above roughly 3 in a week	Rising, often much higher
New creative, same audience	Performance recovers	Little or no change
New audience, same creative	Some recovery	Performance recovers
Audience size	Can be any size	Usually narrow or hyper-local

A word of honesty here: there is no published threshold at which frequency becomes harmful. Anyone who tells you "keep frequency under 2.5" is quoting a rule of thumb, not documentation. What you can observe in your own account is the point at which your cost per result starts climbing as frequency climbs. That number is yours and it differs by business. Find it and write it down.

What does not fix it

Plenty gets sold as the answer here. Some of it is worse than doing nothing.

Fifty variations of the same ad. Agencies sell "creative refresh packages" that produce thirty or fifty assets — the same photo on a blue background, a green background, with the headline moved. This is volume, not variety. If the picture, the promise and the person in the ad are unchanged, a viewer's brain files all fifty as the same ad. Real variation means a different format (static versus video), a different angle (price versus fear versus proof), or a different subject entirely.

Restarting the campaign to "reset the algorithm". Duplicating an ad set and relaunching it throws away accumulated learning and pushes you back into the learning phase, where delivery is unstable until the ad set collects enough optimisation events — Meta's own guidance has long pointed at roughly 50 per ad set per week. You pay for that instability. You have not changed the underlying problem: same audience, same creative.

Cranking the budget up. If you are saturated, doubling spend does not find you double the people. It buys more impressions in the same exhausted pool, raises frequency faster, and burns money quicker. Raising budget is the right move when you have headroom, and the wrong move when you do not. Frequency tells you which you have.

Endlessly broadening the targeting. Removing interest filters can genuinely help, and broad targeting often outperforms narrow. But it is not a cure for a weak offer — it just spreads a weak offer across more people at a higher total cost.

What actually helps

Build a small library before you launch, not after you panic. Three to five genuinely different concepts, shot in one session. A face-to-camera video. A before-and-after. A price-and-terms card. A customer's own photo, used with permission. When performance dips, you rotate rather than scramble.

Rotate on evidence, not on the calendar. Do not swap creative every Monday because someone said so. Swap when your own cost-per-result-versus-frequency curve turns upward.

Exclude people who already converted. If someone has already filled the form or bought, showing them the acquisition ad is wasted money and accelerates apparent saturation. Exclusions are free to set up and almost always neglected.

Accept that some businesses cannot be scaled on paid social alone. If your realistic catchment is 50,000 adults, there is a ceiling, and no amount of creative will lift it. That is the point at which the money is better split — some to ads, some to search visibility that brings in people actively looking, some to the organic account that keeps you familiar between campaigns. Search demand renews itself daily in a way that a fixed audience pool does not.

Run campaigns in bursts. Three weeks on, two weeks off, for a local service business, often produces better blended economics than running flat-out all year. The pool partially refills with new movers, new life stages, new needs.

What to do this week

Open Ads Manager. Set the date range to the last seven days, add the frequency column, then compare with the seven days before that. If frequency is climbing and click-through rate is falling, you now know which problem you have — and the new-creative test will tell you which of the two within a few days.

If you would rather someone else read the numbers and build the rotation, tell us what you are running and what it is costing you. We will tell you whether the problem is the creative, the audience, or something after the click — and we will say so if paid social is not the right place for the next rupee. There is more on ads and measurement across the rest of the blog.

Questions we get asked